By Sean Callahan, Editor-in-Chief
Originally Published November 28, 2023
What truly affects how well students perform in education? Is it a reduced class size or a different school type? Or is it what gender they study with? Is affirmative action helpful? Does school type or competition among schools help enhance academic performance? These were only some of many questions addressed by one of the McKenna’s school’s latest speakers.
On Wednesday, Nov. 15, at 7:30 p.m in the Fred Rogers Center, Dr. Dennis Epple, the Thomas Lord University Professor of Economics at Carnegie Mellon University, gave a talk in which he evaluated different education topics and research regarding school choice and effectiveness. Epple conducts research focused on public economics, with an emphasis on federalism, the economics of education, and urban economics.
Epple began his talk by briefly summarizing various issues and questions of concern in education. Most of his talk and the research he discussed concerned gender, school choice, school test scores, income inequality, and educational vouchers to education.
One prominent topic he discussed was educational vouchers. He affirmed that, according to research, vouchers increase competition between private and public schools, and that parents choose neighborhood based on “perceived school quality”. He also emphasized that income stratification exists across school districts and neighborhoods, using his own research as an example.
“I’ve done a substantial amount of research using the Boston Metropolitan area. School districts and municipal boundaries are coterminous. Whereas in many places, such as New Jersey, municipal boundaries cross.”
He emphasized the “hierarchy of income” that he observed in Boston, which he asserts has remained consistent into the twenty-first century.
Epple also discussed research concerning gender as it relates to academic performance.
“[According to research,] girls perform better in all girls’ schools. On the other hand, boys do better in mixed-gender classes than girls do,” Epple said.
Furthermore, Epple cited the tremendous percentage increase in women obtaining bachelor’s degrees. Whereas 19.2 percent of women had a college degree in 1975 to 1979, compared to men’s 24.6 percent, this skyrocketed to 46.9 percent compared to men’s 39.3 percent from 2015 to 2018.
In terms of race, Epple noted a study that showed a large decrease in high school dropout rates. The graph showed the largest decrease was for black men and women.
“The African American dropout rate has dropped more quickly. When you get to 2021, those rates have dropped to five percent, [from 25 and 30 percent for African American men and women respectively.]”
Epple concluded his lecture with discussion of two approaches to evaluating education research, and how they are applied.
The first approach discussed was regression discontinuity. Epple says this is widely used in education research because “cutoff scores are widely used for admission to educational programs”. This approach is used to obtain unbiased estimates of “education-related interventions”. Epple provided an example of a study concerning a flagship university with an SAT cutoff for university students. By using regression discontinuity, the researcher compared the outcomes of students who barely got into the university and those who fell short on a graph, and discovered that all were “nearly identical in ability but [had] very different opportunities.” A similar example examining students of the same university, found that those above the SAT cutoff point “earned, on average, 9.5 percent more than those just below the cutoff.”
The second approach discussed was random assignment, which, according to Epple, a method of “natural experiments” due to appearing like a lottery. He used the example of the Washington D.C Voucher Experiment, which occurred following Congress’ passage of the District of Columbia School Choice Incentive Act in 2004. The experiment, Epple explained, was implemented by Congress to evaluate the impacts of the program, which provided the first federally funded private school voucher program in the United States. Applicants to private schools were awarded scholarships by lottery (otherwise known as random assignment). The results showed no statistically significant change in student test scores, but they did show a statistically significant increase in high school graduation rates.
The final study Epple discussed in relation to random assignment considered teacher absenteeism. In a 2004 study in India, three unannounced visits were made to 3700 randomly selected schools throughout the country. It was found that the teacher absence rate was 35 percent on average. To mitigate this, in random schools across the country, students were given special cameras by chosen teachers to document when their teachers were present or absent every day. Consistent attendance would be rewarded with extra pay for the teacher. The experiment reduced teacher absenteeism from 35 percent to 14 percent, which Epple called a “creative, relatively inexpensive, and very effective intervention.”
Unrelated to the study, Epple also asserted that “the Constitution of India mandates affirmative action for the two most disadvantaged groups, formerly known as untouchables”, and that the effects were positive. This is in stark contrast, he notes, to the United States, which had a supreme court ruling against affirmative action in college admissions.
The talk ended with several questions from faculty and students. These included an inquiry on Epple’s opinion of the validity of the vouchers and whether the results of the studies could be replicated in other countries aside from India and the United States.